UniFinEval: Towards Unified Evaluation of Financial Multimodal Models across Text, Images and Videos

2026-01-09T10:15:32Z

Multimodal large language models are playing an increasingly significant role in empowering the financial domain, however, the challenges they face, such as multimodal and high-density information and cross-modal multi-hop reasoning, go beyond the evaluation scope of existing multimodal benchmarks. To address this gap, we propose UniFinEval, the first unified multimodal benchmark designed for high-information-density financial environments, covering text, images, and videos. UniFinEval systematically constructs five core financial scenarios grounded in real-world financial systems: Financial Statement Auditing, Company Fundamental Reasoning, Industry Trend Insights, Financial Risk Sensing, and Asset Allocation Analysis. We manually construct a high-quality dataset consisting of 3,767 question-answer pairs in both chinese and english and systematically evaluate 10 mainstream MLLMs under Zero-Shot and CoT settings. Results show that Gemini-3-pro-preview achieves the best overall performance, yet still exhibits a substantial gap compared to financial experts. Further error analysis reveals systematic deficiencies in current models. UniFinEval aims to provide a systematic assessment of MLLMs' capabilities in fine-grained, high-information-density financial environments, thereby enhancing the robustness of MLLMs applications in real-world financial scenarios. Data and code are available at https://github.com/aifinlab/UniFinEval.

Technology Adoption and Network Externalities in Financial Systems: A Spatial-Network Approach

2026-01-09T04:37:54Z

This paper develops a unified framework for analyzing technology adoption in financial networks that incorporates spatial spillovers, network externalities, and their interaction. The framework characterizes adoption dynamics through a master equation whose solution admits a Feynman-Kac representation as expected cumulative adoption pressure along stochastic paths through spatial-network space. From this representation, I derive the Adoption Amplification Factor -- a structural measure of technology leadership that captures the ratio of total system-wide adoption to initial adoption following a localized shock. A Levy jump-diffusion extension with state-dependent jump intensity captures critical mass dynamics: below threshold, adoption evolves through gradual diffusion; above threshold, cascade dynamics accelerate adoption through discrete jumps. Applying the framework to SWIFT gpi adoption among 17 Global Systemically Important Banks, I find strong support for the two-regime characterization. Network-central banks adopt significantly earlier ($ρ= -0.69$, $p = 0.002$), and pre-threshold adopters have significantly higher amplification factors than post-threshold adopters (11.81 versus 7.83, $p = 0.010$). Founding members, representing 29 percent of banks, account for 39 percent of total system amplification -- sufficient to trigger cascade dynamics. Controlling for firm size and network position, CEO age delays adoption by 11-15 days per year.

Vague Knowledge: Information without Transitivity and Partitions

2026-01-07T19:19:37Z

I relax the standard assumptions of transitivity and partition structure in economic models of information to formalize vague knowledge: non-transitive indistinguishability over states. I show that vague knowledge, while failing to partition the state space, remains informative by distinguishing some states from others. Moreover, it can only be faithfully expressed through vague communication with blurred boundaries. My results provide microfoundations for the prevalence of natural language communication and qualitative reasoning in the real world, where knowledge is often vague.

An Algorithmic Framework for Systematic Literature Reviews: A Case Study for Financial Narratives

2026-01-07T10:50:35Z

This paper introduces an algorithmic framework for conducting systematic literature reviews (SLRs), designed to improve efficiency, reproducibility, and selection quality assessment in the literature review process. The proposed method integrates Natural Language Processing (NLP) techniques, clustering algorithms, and interpretability tools to automate and structure the selection and analysis of academic publications. The framework is applied to a case study focused on financial narratives, an emerging area in financial economics that examines how structured accounts of economic events, formed by the convergence of individual interpretations, influence market dynamics and asset prices. Drawing from the Scopus database of peer-reviewed literature, the review highlights research efforts to model financial narratives using various NLP techniques. Results reveal that while advances have been made, the conceptualization of financial narratives remains fragmented, often reduced to sentiment analysis, topic modeling, or their combination, without a unified theoretical framework. The findings underscore the value of more rigorous and dynamic narrative modeling approaches and demonstrate the effectiveness of the proposed algorithmic SLR methodology.

Kladia Liquidity Deflator (KLD): A Debt-Indexed Deflationary Token on XRPL

2026-01-01T15:58:57Z

Kladia Liquidity Deflator (KLD) is an XRPL-based, debt-indexed token whose supply dynamics respond directly to a debt index derived from macroeconomic data sources. The model links indebtedness to deterministic adjustments in issuance, burns, and escrow release caps, creating a rule-based deflationary mechanism that strengthens as debt rises. With a fixed maximum supply of 10 billion KLD, the mechanism is implemented through XRPL oracles and governance. Escrow locking depends on the TokenEscrow amendment; until it is active network-wide, allocations will be secured in a multi-signature vault with published rules and public monitoring. KLD provides a transparent and mathematically grounded framework for a macro-responsive digital asset.

General Equilibrium Amplification and Crisis Vulnerability: Cross-Crisis Evidence from Global Banks

2026-01-01T05:17:08Z

This paper develops a continuous framework for analyzing financial contagion that incorporates both geographic proximity and interbank network linkages. The framework characterizes stress propagation through a master equation whose solution admits a Feynman-Kac representation as expected cumulative stress along stochastic paths through spatial-network space. From this representation, I derive the General Equilibrium Amplification Factor -- a structural measure of systemic importance that captures the ratio of total system-wide effects to direct effects following a localized shock. The amplification factor decomposes naturally into spatial, network, and interaction components, revealing which transmission channels contribute most to each institution's systemic importance. The framework nests discrete cascade models as a limiting case when jump intensity becomes infinite above default thresholds, clarifying that continuous and discrete approaches describe different regimes of the same phenomenon. Empirical validation using 38 global banks across the 2008 financial crisis and COVID-19 pandemic demonstrates that the amplification factor correctly identifies systemically important institutions (Pearson correlation $ρ= -0.450$, $p = 0.080$ between amplification factor and crisis drawdowns) and predicts crisis outcomes out-of-sample ($ρ= -0.352$ for COVID-19). Robustness analysis using cumulative abnormal returns -- a measure more directly connected to the Feynman-Kac integral -- strengthens these findings ($ρ= -0.512$, $p = 0.042$). Time-series analysis confirms that average pairwise bank correlations track macroeconomic stress indicators ($ρ= 0.265$ with VIX, $p < 0.001$). Comparing the two crises reveals that COVID-19 produced a sharper correlation spike (+93%) despite smaller equity losses, reflecting different contagion dynamics for exogenous versus endogenous shocks.

SoK: Stablecoins in Retail Payments

2026-01-01T04:06:44Z

Stablecoins have emerged as a rapidly growing digital payment instrument, raising the question of whether blockchain-based settlement can function as a substitute for incumbent card networks in retail payments. This Systematization of Knowledge (SoK) provides a systematic comparison between stablecoin payment arrangements and card networks by situating both within a unified analytical framework. We first map their respective payment infrastructures, participant roles, and transaction lifecycles, highlighting fundamental differences in how authorization, settlement, and recourse are organized. Building on this mapping, we introduce the CLEAR framework, which evaluates retail payment systems across five dimensions: cost, legality, experience, architecture, and reach. Our analysis shows that stablecoins deliver efficient, continuous, and programmable settlement, often compressing rail-level merchant fees and enabling 24/7 value transfer. However, these advantages are accompanied by an inversion of the traditional pricing and risk-allocation structure. Card networks internalize consumer-side frictions through subsidies, standardized liability rules, and post-transaction recourse, thereby supporting mass-market adoption. Stablecoin arrangements, by contrast, externalize transaction fees, error prevention, and dispute resolution to users, intermediaries, and courts, resulting in weaker consumer protection, higher cognitive burden at the point of interaction, and fragmented acceptance. Accordingly, stablecoins exhibit a conditional comparative advantage in closed-loop environments, cross-border corridors, and high-friction payment contexts, but remain structurally disadvantaged as open-loop retail payment instruments.

Who sets the range? Funding mechanics and 4h context in crypto markets

2025-12-31T00:19:59Z

Financial markets often appear chaotic, yet ranges are rarely accidental. They emerge from structured interactions between market context and capital conditions. The four-hour timeframe provides a critical lens for observing this equilibrium zone where institutional positioning, leveraged exposure, and liquidity management converge. Funding mechanisms, especially in perpetual futures, act as disciplinary forces that regulate trader behavior, impose economic costs, and shape directional commitment. When funding aligns with the prevailing 4H context, price expansion becomes possible; when it diverges, compression and range-bound behavior dominate. Ranges therefore represent controlled balance rather than indecision, reflecting strategic positioning by informed participants. Understanding how 4H context and funding operate as market governors is essential for interpreting cryptocurrency price action as a rational, power-mediated process.

Does Peer-Reviewed Research Help Predict Stock Returns?

2025-12-29T20:02:01Z

Mining 29,000 accounting ratios for t-statistics $> 2.0$ leads to cross-sectional return predictability similar to the peer review process. For both, $\approx50\%$ of predictability remains after the original sample periods. This finding holds for many categories of research, including research with risk or equilibrium foundations. Only research agnostic about the theoretical explanation for predictability shows signs of outperformance. Our results imply that inferences about post-sample performance depend little on whether the predictor is peer-reviewed or data mined. They also have implications for the importance of empirical vs theoretical evidence, investors' learning from academic research, and the effectiveness of data mining.

The Nonstationarity-Complexity Tradeoff in Return Prediction

2025-12-29T16:49:19Z

We investigate machine learning models for stock return prediction in non-stationary environments, revealing a fundamental nonstationarity-complexity tradeoff: complex models reduce misspecification error but require longer training windows that introduce stronger non-stationarity. We resolve this tension with a novel model selection method that jointly optimizes model class and training window size using a tournament procedure that adaptively evaluates candidates on non-stationary validation data. Our theoretical analysis demonstrates that this approach balances misspecification error, estimation variance, and non-stationarity, performing close to the best model in hindsight. Applying our method to 17 industry portfolio returns, we consistently outperform standard rolling-window benchmarks, improving out-of-sample $R^2$ by 14-23% on average. During NBER-designated recessions, improvements are substantial: our method achieves positive $R^2$ during the Gulf War recession while benchmarks are negative, and improves $R^2$ in absolute terms by at least 80bps during the 2001 recession as well as superior performance during the 2008 Financial Crisis. Economically, a trading strategy based on our selected model generates 31% higher cumulative returns averaged across the industries.

Deep Learning for Art Market Valuation

2025-12-28T21:04:09Z

We study how deep learning can improve valuation in the art market by incorporating the visual content of artworks into predictive models. Using a large repeated-sales dataset from major auction houses, we benchmark classical hedonic regressions and tree-based methods against modern deep architectures, including multi-modal models that fuse tabular and image data. We find that while artist identity and prior transaction history dominate overall predictive power, visual embeddings provide a distinct and economically meaningful contribution for fresh-to-market works where historical anchors are absent. Interpretability analyses using Grad-CAM and embedding visualizations show that models attend to compositional and stylistic cues. Our findings demonstrate that multi-modal deep learning delivers significant value precisely when valuation is hardest, namely first-time sales, and thus offers new insights for both academic research and practice in art market valuation.

Mean-Field Price Formation on Trees with a Network of Relative Performance Concerns

2025-12-25T10:50:09Z

Financial firms and institutional investors are routinely evaluated based on their performance relative to their peers. These relative performance concerns significantly influence risk-taking behavior and market dynamics. While the literature studying Nash equilibrium under such relative performance competitions is extensive, its effect on asset price formation remains largely unexplored. This paper investigates mean-field equilibrium price formation of a single risky stock in a discrete-time market where agents exhibit exponential utility and relative performance concerns. Unlike existing literature that typically treats asset prices as exogenous, we impose a market-clearing condition to determine the price dynamics endogenously within a relative performance equilibrium. Using a binomial tree framework, we establish the existence and uniqueness of the market-clearing mean-field equilibrium in both single- and multi-population settings. Finally, we provide illustrative numerical examples demonstrating the equilibrium price distributions and agents' optimal position sizes.

Behavioral Machine Learning? Regularization and Forecast Bias

2025-12-23T18:23:14Z

Standard forecast efficiency tests interpret violations as evidence of behavioral bias. We show theoretically and empirically that rational forecasters using optimal regularization systematically violate these tests. Machine learning forecasts show near zero bias at one year horizon, but strong overreaction at two years, consistent with predictions from a model of regularization and measurement noise. We provide three complementary tests: experimental variation in regularization parameters, cross-sectional heterogeneity in firm signal quality, and quasi-experimental evidence from ML adoption around 2013. Technically trained analysts shift sharply toward overreaction post-2013. Our findings suggest reported violations may reflect statistical sophistication rather than cognitive failure.

Diagram-to-Circuit QNLP for Financial Sentiment Analysis

2025-12-23T04:54:52Z

We study a \emph{QDisCoCirc}-inspired, chunked diagram-to-circuit quantum natural language processing (QNLP) model for three-class sentiment classification of financial texts. In our classical simulations, we keep the Hilbert-space dimension manageable by decomposing each sentence into short contiguous chunks. Each chunk is mapped to a shallow quantum circuit, and the resulting Bloch vectors are used as a sequence of quantum tokens. Simple averaging of chunk vectors ignores word order and syntactic roles. We therefore add a small Transformer encoder over the raw Bloch-vector sequence and attach a CCG-based type embedding to each chunk. This hybrid design preserves physically interpretable semantic axes of quantum tokens while allowing the classical side to model word order and long-range dependencies. The sequence model improves test macro-F1 over the averaging baseline and chunk-level attribution further shows that evidential mass concentrates on a small number of chunks, that type embeddings are used more reliably for correctly predicted sentences. For real-world quantum language processing applications in finance, future key challenges include circuit designs that avoid chunking and the design of inter-chunk fusion layers.

Cryptocurrencies and Interest Rates: Inferring Yield Curves in a Bondless Market

2025-12-17T10:21:00Z

In traditional financial markets, yield curves are widely available for countries (and, by extension, currencies), financial institutions, and large corporates. These curves are used to calibrate stochastic interest rate models, discount future cash flows, and price financial products. Yield curves, however, can be readily computed only because of the current size and structure of bond markets. In cryptocurrency markets, where fixed-rate lending and bonds are almost nonexistent as of early 2025, the yield curve associated with each currency must be estimated by other means. In this paper, we show how mathematical tools can be used to construct yield curves for cryptocurrencies by leveraging data from the highly developed markets for cryptocurrency derivatives.